TraffickCam is a product of Imaging for Good, a 501(c)3 dedicated to the creation of technology for social impact.
The research and development of the TraffickCam application and image search is led by Dr. Abby Stylianou at Saint Louis University, Dr. Robert Pless at George Washington University, Dr. Richard Souvenir at Temple University, and Dr. Nathan Jacobs at Washington University in St. Louis. This team of researchers has applied their collective decades of experience in computer vision and machine learning to develop state of the art approaches to recognizing hotels in images of victims of sex trafficking.
The research and development of the TraffickCam system has been supported by a variety of sources, including the Congregation of the Sisters of Saint Joseph, the National Institute of Justice, and the Laboratory of Analytic Sciences, and has led to an assortment of peer reviewed research publications:
- Traffickcam: Crowdsourced and computer vision based approaches to fighting sex trafficking, AIPR 2017
- Visualizing deep similarity networks, WACV 2019
- Hotels-50k: A global hotel recognition dataset, AAAI 2020
- Improved embeddings with easy positive triplet mining, WACV 2020
- Hard negative examples are hard, but useful, ECCV 2020
- Visualizing paired image similarity in transformer networks, WACV 2022
- Exploring CLIP for Real World, Text-based Image Retrieval, AIPR 2023
- Hotel Recognition Using Object Ensembles, AIPR 2023
Our History
TraffickCam was first created in 2015 as a collaboration between the Exchange Initiative and Drs. Stylianou, Pless and Souvenir.
The Exchange Initiative was created by Nix Conference & Meeting Management to empower individuals and organizations with real resources to help end sex trafficking. Nix Conference & Meeting Management is one of just 13 U.S. companies and 43 worldwide honored as a 2014 Top Member by the internationally recognized Tourism Child-Protection Code of Conduct (TheCode.org) for their exceptional work to integrate child protection practices into their business.